The asynchronous interconnection of large grids transforms the power system from a large grid operating synchronously to a small grid operating independently, which enhances the sensitivity to grid frequency deviation perturbations. Automatic Generation Control (AGC) is an important way to regulate the deviation of the modern grid parameters. The serious deviation of the grid parameters will lead to the occurrence of power accidents, and even small-scale frequency fluctuations will also pose a threat to the AGC control. In order to enhance the stability and security of the power grid, a nonlinear collaborative control method for distributed photovoltaic energy storage power station was designed. The constraints of the AGC control model were set by deeply analyzing the data composition structure of the grid and the load characteristics of the power station after asynchronous networking. The multi-objective control equations of the asynchronously networked grid were established according to the deviation amount of control objectives of different mini-grids. This method combines the control law of space power station system and realizes the nonlinear collaborative control of distributed photovoltaic energy storage power stations through the calculation of optimal deviation control variables. The experimental results show that the AGC control of asynchronous networked grids using this method can lead to a smooth trend of active power output at grid connection, and it also can effectively control the charging and discharging power in the charging state of the power station. Moreover, the method improves the response speed of the system significantly and reduces the transient response control error, thereby enhancing the collaborative control effect significantly.